Abstract: The LRU-like algorithm is proposed for the block management of augmented cache scheme. Based on the LRU-like algorithm, a kind of augmented cache named least-recently-used blocks filter cache (LBF cache) is designed. After the simulation, the performance of LBF cache shows better than some augmented caches with similar architectures (such as victim cache and assist cache), and also better than the traditional direct-mapping cache with double size.

Key words: Cache performance, Hit ratio, LRU algorithm

摘要： 提出了一个用于扩充高缓块管理的近似LRU算法。利用该算法，设计了一个可过滤LRU数据块的扩充高缓方案——LRU块过滤高缓(LBF高缓)。仿真结果显示，LBF高缓的性能优于类似结构的扩充高缓(如牺牲高缓和辅助高缓)，与具有2倍容量的直接映像高缓相比性能有所提高。

关键词: 高缓性能, 命中率, LRU算法

CLC Number: 

• TP333